First published 2011

by Psychology Press
27 Church Road, Hove, East Sussex BN3 2FA
Simultaneously published in the USA and Canada
by Psychology Press
270 Madison Avenue, New York, NY 10016
Psychology Press is an imprint of the Taylor & Francis Group, an
Informa business
Copyright 2011 Psychology Press
Typeset in Times by ReneCatch Limited, Bungay, Suolk
Printed and bound in Great Britain by
TJ International Ltd, Padstow, Cornwall
Cover design by Andrew Ward
All rights reserved. No part of this book may be reprinted or
reproduced or utilized in any form or by any electronic,
mechanical, or other means, now known or hereafter
invented, including photocopying and recording, or in any
information storage or retrieval system, without permission in
writing from the publishers.
This publication has been produced with paper manufactured to strict
environmental standards and with pulp derived from sustainable forests.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
The science of reason : a Festschrift for Jonathan St. B. T. Evans /
edited by Ken Manktelow, David Over, and Shira Elqayam.
p. cm.
Includes bibliographical references and index.
1. Reasoning (Psychology). 2. Thought and thinking. 3. Problem
solving. I. Evans, Jonathan St. B. T., 1948. II. Manktelow,
K. I., 1952. III. Over, D. E., 1946. IV. Elqayam, Shira.
BF442.S25 2010
153.43dc22
2010002556
ISBN: 978-1-84872-015-2 (hbk)

http://www.psypress.com/the-science-of-reason-9781848720152

Contents

List of gures
List of tables
List of contributors
1 Paradigms shift: Jonathan Evans and the science of reason

ix
xi
xiii
1

KEN MANKTELOW, DAVID OVER, AND SHIRA ELQAYAM

PART I

Thinking and reasoning: Psychological approaches

2 Solving natural syllogisms

17
19

GUY POLITZER

3 Probability evaluations, expectations,

and choices

37

VITTORIO GIROTTO AND MICHEL GONZALEZ

4 Arguments in mind

53

DAVID W. GREEN

5 Facilitation and analogical transfer on a

hypothetico-deductive reasoning task

63

CYNTHIA S. KOENIG AND RICHARD A. GRIGGS

PART II

91

If
6 Conditionals and disjunctions
KLAUS OBERAUER, SONJA GEIGER, AND
KATRIN FISCHER

http://www.psypress.com/the-science-of-reason-9781848720152

93

vi

Contents

7 The truth about conditionals

119

PHILIP N. JOHNSON-LAIRD

8 Uncertain deductive reasoning

145

NIKI PFEIFER AND GERNOT D. KLEITER

9 Thinking before you decide on the selection task:

Matching bias requires analytical reasoning

167

SIMON J. HANDLEY, STEPHEN E. NEWSTEAD, AND

HELEN NEILENS

10 Dual processes in the development of the understanding

of conditionals

191

PIERRE BARROUILLET AND CAROLINE GAUFFROY

PART III

Dual processes and beyond

217

11 Dual-process theories of thinking

219

PAOLO LEGRENZI

12 A conjunction of fallacies: What dierent types of

causal conjunction error reveal about dual processes
for thinking

237

AIDAN FEENEY AND AIME CRISP

13 Scalar implicature: Inference, convention, and

dual processes

259

KEITH FRANKISH AND MARIA KASMIRLI

14 The dynamics of reasoning: Chronometric analysis and

dual-process theories

283

LINDEN J. BALL

15 Methodological and theoretical issues in belief bias:

Implications for dual-process theories

309

VALERIE A. THOMPSON, STEPHEN E. NEWSTEAD, AND

NICOLA J. MORLEY

16 Dual systems and dual processes but a single function

MIKE OAKSFORD AND NICK CHATER

http://www.psypress.com/the-science-of-reason-9781848720152

339

Contents

vii

PART IV

Rationality and reasoning

353

17 Individual dierences as essential components of

heuristics and biases research

355

KEITH E. STANOVICH, RICHARD F. WEST, AND

MAGGIE E. TOPLAK

18 Grounded rationality: A relativist framework for

normative rationality

397

SHIRA ELQAYAM

PART V

Perspectives on thinking and reasoning

421

19 The psychology of reasoning: Reections on four

decades of research

423

JONATHAN ST B. T. EVANS

Author index
Subject index

http://www.psypress.com/the-science-of-reason-9781848720152

445
455

18 Grounded rationality
A relativist framework for
normative rationality
Shira Elqayam
De Montfort University

An old Jewish story tells about two litigants appearing before the towns
rabbi. The rst litigant having presented his case, the rabbi tells him: You are
right. Then the second litigant presents his case, and the rabbi tells him, too:
You are right. After the litigants had gone, the rabbis wife reproaches him:
How can they both be right? The rabbi listens to her attentively, and says: You
know what? You are right, too!
Stories like this reect the deep unease that some of us feel about relativism, the you are right too view, the idea that judgement should be done in
context. How can everyone be right? And when rationality is the issue, relativism seems a particularly bad idea. It seems that one cannot be relatively
rational: the question of rationality is an absolute one. As Stein (1996) memorably put it: prima facie, relativism seems implausible with respect to reasoning. It seems crazy to say that reasoning in accordance with principles
based on rules of logic is a good thing for some people but not for others
(Stein, 1996, p. 32; more on Steins critique later). Oaksford and Chater
recently voiced a similar objection, saying there is a strong intuition [. . .]
that there is some absolute sense in which some reasoning or decision-making
is good, and other reasoning and decision-making is bad (Oaksford &
Chater, 2007, pp. 2425).
In this chapter, I address the issue of relativism in rationality.1 Rationality
seems a suitable topic for a book in honour of Jonathan Evanss sixtieth
birthday: Jonathans contribution to the eld of human rationality (much of
it with David Over) has shaped the way many of us think about rationality. In
their seminal Rationality and Reasoning (1996), Evans and Over famously
defended a distinction between rationality1, the pragmatic, instrumental sort
of rationality involved in obtaining goals, and rationality2, implied by explicit
normative rule following. This distinction is pertinent to the present discussion: relativism has a dierent meaning in the context of instrumental
rationality and normative rationality. In instrumental rationality, a relativist
viewpoint means that goals are relative to the individual, or, even more
strongly, that cognitive tools for obtaining these goals are relative too. This is
the approach that Evans and Over take regarding instrumental rationality:
rationality1 is personal and relative to the individual (p. 4), focusing on

http://www.psypress.com/the-science-of-reason-9781848720152

398

Elqayam

what our individual goals are, and whether we are reasoning or acting in a
way that is generally reliable for achieving these (p. 7). It is explicitly relative.
Similarly, Oaksford and Chater have recently defended the goal relativity of
rational analysis (Oaksford & Chater, 2009, p. 100).
In comparison, in the context of normative rationality, relativism means
that normative systems are relative to an individual in a particular context
and should be judged in that context. That is, if an individual explicitly follows
her own relevant normative system and conforms to its rules, then from a
relativist angle this is both necessary and sucient to make her normatively
rational (rational2 in Evans and Overs terms). A relativist viewpoint means
that the normative system followed does not have to be objectively justied or
impersonal, although it does have to be subjectively and contextually justied. The issue seems orthogonal to the question of relativism in instrumental
rationality: one can grant that goals are relative to the individual, that reliability
of various cognitive tools for achieving them is relative, and yet leave open the
issue whether normative systems are relative, or even opt for an absolutist view
of the latter. While Evans and Over (1996) do not explicitly discuss the issue,
they seem to adopt the latter view: rationality2 is presented as impersonal.
In this chapter, I will take one step further a view I have recently defended
in collaboration with Jonathan (Elqayam & Evans, 2010), and will argue that,
if psychologists wish to preserve a notion of normative rationality, this can
best be done within a relativist framework. I will then examine a possible
framework for such a relativist conception of normative rationality and
defend it. I will argue that normative judgements in psychology run the risk of
a problematic inference, whose solution requires a relativist viewpoint. I will
also argue that the same principle that underlies bounded rationality, the
ought implies can principle, means that normative judgement has to be
taken in context. I will start by demonstrating how some of the arguments
against normativism undermine an absolutist framework of normative
rationality and that a relativist framework is a viable potential solution. Moving on to describing the relativist framework, I will start by sketching the way
normative relativism changes the normative research question. The following
section examines in some detail the way relativism would work for normative
rationality. I will propose and defend a conception of grounded rationality,
which is a relativist version of bounded rationality. Next examined is a major
question that relativism and grounded rationality have to deal with, the role
of universal constraints. Next, I examine the role of explicit processing.
Finally, I will look in some detail at possible objections to the proposed
relativist framework.

Relativism, normativism, and inferring the ought from the is

Notwithstanding what seems like the counterintuitive character of relativism
in human thought, relativism does acquire greater plausibility in historical
context. Plurality of formal systems is a matter of record in the history of

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

399

science. Like any domain of knowledge, formal systems develop and change
over time and space (see Olson & Torrance, 1996). Consider the history of
logic, mathematics and statistics the leading candidates for normative systems. The simple observation is that they keep changing. The logic of, say,
twenty-rst century textbooks, is hardly what medieval students learned in
the trivium. The highly sophisticated methods that Chinese mathematicians,
for example, used to obtain results comparable in accuracy to those of the
Greek mathematicians, were entirely dierent, involving no use of axioms
(Lloyd, 1996). Contemporary formal systems vary too think about manyvalued logics (Gottwald, 2001; Rescher, 1969), or non-Euclidean geometries
(Greenberg, 2008).
Add to this the accumulating evidence for cultural and individual dierences in the psychology of reasoning and judgement and decision making.
Dierent people do seem to think dierently, depending on cultural (Buchtel
& Norenzayan, 2009; Nisbett, Peng, Choi, & Norenzayan, 2001) and individual dierences in style (Epstein, 1994; Epstein, Pacini, Denes-Raj, & Heier,
1996), ability and cognitive motivation (Stanovich, 1999, 2004; Stanovich &
West, 2000). Even what people count as knowledge seems to be culture sensitive (Weinberg, Nichols, & Stich, 2006; also see Nichols, 2004). The case for
descriptive pluralism, then (in the term coined by Stich, 1990), seems beyond
dispute. And from descriptive pluralism to normative relativism seems but
one easy step.
So here is the paradox: On the one hand, this intuitive prima facie
implausibility (in Steins terms) of normative relativism; on the other the
historical, geographical, and psychological evidence for descriptive pluralism.
Of course, a step from is to ought is never easy! One can accept the
evidence for descriptive cultural pluralism and yet defend an absolutist normative view. However, this very gap between the is and the ought, the factual
and the evaluative, provides us with a very good reason to embrace a relativist
approach to human thought.
Recently, Elqayam and Evans (2010) identied and argued against normativism as a dominant paradigm in the psychology of human thinking. Normativism was dened as the view that (N1) human thinking reects some
kind of formal system S; and (N2) should be judged against S as a normative
system. A major part of the argument was based on the descriptive pluralism
of formal systems. Several authors (Evans, 1993; Stanovich, 1999) have
pointed out a problem of arbitration (dubbed the normative system problem and the inappropriate norm argument respectively), when more than
one normative system ts in with a particular empirical corpus. Elqayam and
Evans (2010) pointed out that the problem of arbitration is particularly acute
for normativism, because normative systems assume an ought component
that is absent from computational (in Marrs terms) or competence (in
Chomskys terms) level theories (also see Elqayam, 2007; Evans, 2009a;
Schroyens, 2009). Elqayam and Evans argued that, while supporting is
theory with is evidence is fairly straightforward, supporting an ought

http://www.psypress.com/the-science-of-reason-9781848720152

400

Elqayam

theory with is evidence seems to invite an isought (Hume, 2000) or

naturalistic (Moore, 1903) inference: where the premises are purely factual,
inferring evaluative, deontic conclusions has been argued to be fallacious
(Hudson, 1969; Schurz, 1997). A notorious example of the naturalistic fallacy in fact, the one that prompted Moores criticism is Spencers Social
Darwinism (Hofstadter, 1955; although see Weinstein, 2002), which converts
Darwinian adaptation mechanisms (is) into a form of extreme capitalism
captured by the catchphrase survival of the ttest, and presented as a
moral argument (ought). Since any empirical solution to the arbitration
problem requires inferring an ought (normative system) from an is
(empirical evidence), normativism is problematic for psychology; Elqayam
and Evans concluded, therefore, with an outline of a descriptivist agenda in
the psychological study of human thinking.
It is far from clear, however, that inferring the ought from the is is
indeed a fallacy. A tradition in meta-ethics endeavours to clear isought
inference: the core argument is that the separation of factual from evaluative is articial, and that many acts of communication (concepts, speech
acts) in ordinary language are both (e.g., Frankena, 1939; Searle, 1964;
Williams, 1985). For example, if Jones promises to pay Smith 5, then
Jones ought to pay Smith 5; no fallacy is involved, since promising as a
deontic speech act already incorporates an ought (Searle, 1964). Elqayam
and Evans acknowledge this; but, they argue, here is the rub: these amalgam concepts (thick concepts, to borrow a term from Williams, 1985)
typically depend on social construction. It is social consensus that gives
them the evaluative, deontic meaning, and social consensus is relative to the
society. The amalgam solution seems to be a package deal: if we use it to
support normative judgement, such judgement would have to be in context.
To save normative rationality from the isought problem, then, one has
to opt for a solution that depends on a type of relativism alien to
normativism.
The amalgam solution, then, cannot save normativism, but it can and it
does support relativism. Normative evaluation free from the isought problem is only possible within a relativistic framework. Stich (1990) has already
mounted a comprehensive support of relativism in instrumental rationality, a
view shared, as mentioned, by Evans and Over (1996) and to some extent by
Oaksford and Chater (2009). However, the amalgam solution to the is
ought problem implies that relativism is mandatory in order to preserve
some notion of normative rationality in psychology. The price of saving
normative rationality, then, is giving up on comparison against an absolute
evaluative norm. But this changes everything: many of the research questions in the rationality debate depend on the normativist paradigm and
share its presupposition. In the following sections, I will pick up where
Elqayam and Evans (2010) have stopped, and examine in some detail the
way a relativist paradigm changes the face of research where normative
rationality is concerned.

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

401

Relativism: The research question

Once the normativist paradigm is discarded, then the psychology of reasoning
and judgement has two main options. The rst one, advocated by Elqayam
and Evans (2010), is a purely descriptivist agenda, giving up any normative
considerations. However, it seems that relativism allows a second option:
retain a modicum of normative judgement, provided this is done in relativist
context. The relativist option, however, is not that far dierent from a purely
descriptivist agenda. I will take this issue up again in the nal section, but
before that, I should examine in some detail what normative relativism means.
So how does relativism change the agenda of normative rationality? I will
start with the basics: the research question is dierent. The typical normativist,
absolutist research question is Qna :
Qna. Is formal system S normative?
With relativism, Qna becomes too underspecied, and is no longer even
meaningful; it would be replaced by Qnr :
Qnr. Is formal system S normative for agent A in epistemic context E?
Where epistemic context is both diachronic and synchronic, and sensitive to
culture, society, geography, ecology, as well as the individual: anything that
can have an eect on ones acceptance of a particular normative system.
Thus, both personal and collective inuences are taken into account in Qnr.
Note that relativism obviates the arbitration problem or at least attenuates it
considerably: Dierent normative systems, S1 and S2, can be normative at the
same time, if applied to dierent agents A1 or A2, or dierent epistemic
contexts E1 or E2, or both.
I should also note that normative relativism is dierent from the instrumental or pragmatic sort, which acquired prominence in Stichs seminal
defence of relativism (Stich, 1990), and which can be summarized as:
Qpr. Is [formal] system S useful (ecient, conducive to goal attainment)
for agent A in epistemic context E?
One of the main dierences in epistemic context is the one between analytic,
decontextualizing thinking style and holistic, contextualizing ones. There
are cultures in which such contextual holistic thinking is the normative,
trained response (Buchtel & Norenzayan, 2009; Choi, Koo, & Choi, 2007;
Koo & Choi, 2005). The cultural dierences literature tends to focus on
the Eastern/Western dichotomy, but the dierences seem to be a matter of
training rather than geography: for example, responses of Korean students
diered for Oriental medicine versus psychology majors, the latter responding similarly to Western students (Koo & Choi, 2005). I will therefore prefer

http://www.psypress.com/the-science-of-reason-9781848720152

402

Elqayam

the terms contextualizing versus decontextualizing cultures or epistemic

contexts.
Just like normativism, relativism can vary, and seems to be a matter of
degree rather than a dichotomy. Elqayam and Evans (2008) argued that both
N1 (human thinking reects normative system S) and N2 (S is the normative
rational standard) are better conceived as vectors rather than dichotomies:
N2, in particular, is sensitive to questions such as the a priori status of the
normative system, whether conforming to it is considered both necessary and
sucient to normative rationality or only necessary, and whether the same
normative system holds universally or only for each specic task. Relativism,
as the mirror image of N2, is sensitive to the same factors. At its very extreme,
each unique combination of individual in epistemic context should be tested
for rational norms and conforming to them; the more one allows room for
universal norms and behaviours, the weaker the relativism. In the next section, I will elaborate more on the role of universal and particular factors in a
possible relativist framework.

Bounded rationality and grounded rationality:

A relativist framework
In the previous sections, I have defended a relativist conceptualization of
normative rationality by, following Elqayam and Evans (2008), demonstrating it to be a viable solution to the isought dilemma. To some of the readers
of this chapter, though, this defence may seem rather ad hoc, and relativism
too high a price to pay for some quaint nicety. Now that we have examined
the research question, however, we can see that there are theoretical benets
to be reaped by adopting a relativist position that go beyond solution to the
arbitration problem. First, normativism in its absolutist form runs a substantial risk of cultural blindness (Norenzayan & Heine, 2005), or, in the more
general term coined by Stich (1990), epistemic chauvinism: the tendency
to generalize individual intuitions to universal norms. While epistemic chauvinism is always a risk for normative rationality in any form, the risk is
considerably lower with relativism. It may be impossible to take epistemic
context fully into account: after all, each situation is a unique combination of
individual, cultural and historical context. But overgeneralization is a higher
risk if all this is completely ignored.
More importantly, relativism adds a signicant dimension to the traditional concerns of bounded rationality. Simons (1957, 1982, 1983) model of
bounded rationality refers mainly to the way that human rationality is constrained by what Cherniak (1986) calls the nitary predicament; that is,
physical and cognitive limitations on processing. The fact that humans do not
live forever, that our brain capacity is nite, dictates that only tractable computations could count as rational. The underlying rationale is the age old
ought implies can attributed to Kant (1932/1787).
In addition to bounded rationality, once epistemic context is taken into

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

403

account, it also makes sense to talk about what I will term grounded rationality: that is, normative as well as instrumental rationality that is grounded in
the individuals epistemic context and everything that goes into it: species
specic cognition, personal ability and cognitive style, history, culture, society, and ecology.2 Grounded rationality does not replace bounded rationality;
they are conceptually distinct ideas, contextual versus universalist respectively, albeit based on a similar ought implies can principle. Note that
bounded rationality itself is a relativist kind of constraint (Nickerson, 2008):
it binds rationality to the relative context of human nitary considerations.
However, from an angle of normative relativism, bounded rationality is just
the beginning.
Accepting grounded rationality considerations means that the psychology
of human thinking needs to consider seriously where contextual epistemic
boundaries lie. Cultural dierences are perhaps the prototypical (albeit not
the only) source of epistemic context. They are certainly more accessible for
psychological research: one can hardly time travel to the fourteenth century
and study William of Ockhams contemporaries under controlled experimental conditions. Cross-cultural studies,3 therefore, seem to be the best
bet for analysing the eects of epistemic context. It is easy to forget that
within-culture studies represent a biased sample of human behaviour what
Norenzayan and Heine (2005) call the restricted database of psychology.
This oft-neglected truism has far-reaching implications for bounded rationality as a universalist concept: it is easy to mistake cultural constraints for
universal cognitive limitations, when the sample is biased towards one culture. To avoid this error, we need to understand where bounded rationality
ends and grounded rationality begins.
There is a precedent in cognitive science. Much of the ground I am trying
to cover has already been covered in linguistics, where universalism has
been a signicant issue for many decades. In a way, cross-cultural studies of
human thinking are now where linguistics was about half a century ago,
perhaps because the easily accessible evidence is dierent in each case. In
language, the empirical facts of linguistic diversity are blindingly obvious:
dierent languages have dierent lexicons, dierent grammars. It took the
Chomskyan revolution for linguistics to focus on language universals, to
look for constraints on language that go beyond those imposed by particular
languages (although see Harris, 1980, for a social constructionist critique of
this approach; and Evans & Levinson, 2009, for a recent cognitive science
critique).
In contrast, cross-cultural dierences in thought may not be all that obvious or accessible, and it is only in the last two decades or so that psychologists
have started to have some inkling of their extent, mainly thanks to the eorts
of Nisbett, Norenzayan, and their collaborators (Norenzayan & Heine,
2005; Buchtel & Norenzayan, 2009; Nisbett et al., 2001). It is even more
recently that authors have started comparing the role of universal and cultural constraints on thinking, in particular within a dual-process framework

http://www.psypress.com/the-science-of-reason-9781848720152

404

Elqayam

(Buchtel & Norenzayan, 2009; Norenzayan & Heine, 2005; Yama, Nishioka,
Horishita, Kawasaki, & Taniguchi, 2007). There is no guarantee that thinking is indeed analogous to language and that the role of universal and local
constraints can be protably transferred from the study of the latter to the
study of the former; however, it seems a useful starting point, at least as a
work postulate. The next section looks into this analogy in more detail.
Grounded rationality and the question of cognitive universals
The analogy between linguistic and thinking universals and their implications
for bounded rationality is illustrated in Figure 18.1. The rst part of the
gure, 18.1(a), centres on the role of linguistic universals. To explain it, I will
draw on a classic example of linguistic universals: the order in which each
language prefers to arrange the subject (S), verb (V), and object (O) in a
normal (unmarked) sentence (Comrie, 1981; Greenberg, 1963; Hawkins,
1983). A brief consideration shows that with three components, potentially
there can be six permutations of word order in any language: SVO, SOV,
VSO, VOS, OVS, and OSV. The outermost of the concentric circles in
Figure 18.1(a) represents this possibility. For word order, the possibilities
embodied in all languages, natural as well as formal (i.e., logic), would include
these six permutations. However, out of these six, only three are prevalent in
known natural human languages: SOV (subjectobjectverb), SVO, and VSO
(in that order). The other orders are much rarer. This would be captured in
the intermediate circle in 18.1(a), which represents linguistic universals
in natural human languages. The classic explanation that goes back to
Greenberg (1963) is that linguistic universals in general and word order universals in particular are a product of constraints on processing: in all the
prevalent orders, subject precedes the object (also see Evans & Levinson,
2009). Lastly, particular languages, represented in the innermost circle, tend
to adopt specic orders as the unmarked, default word order: English is an
SVO language, Arabic is a VSO language, and so on.
The potential possibilities space, then, is narrowed down by constraints on
processing that determine language universals, and that is further narrowed
down in particular languages. Note that in this case as in most cases, neither
type of constraint is entirely rigid: other word orders exist in world languages even if they are far less common (although whether they all exist is
moot); some languages do not have any preferred order; and in languages
with preferred order, the constraints only apply to the unmarked, normal
sentence structure. Dierent word order may be used, mainly for emphatic
purposes. (Compare, for example, this pair of SVO and OSV English sentences: The police detained the suspect. The witnesses, they released after
questioning.) How far universality extends and what role linguistic diversity
has is moot. Evans and Levinson (2009) have recently defended a strong
diversity-based approach to universals, maintaining that diversity exceeds
universality, and that whatever universals do exist reect cognitive rather than

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

405

Figure 18.1 (a) Linguistic universals, (b) human thought, and (c) bounded and
grounded rationality.

http://www.psypress.com/the-science-of-reason-9781848720152

406

Elqayam

linguistic constraints. What is not moot, however, is that both universality

and diversity play a role, and that at least some linguistic universals are a
function of cognitive universals.
The analogy to human thought and rationality can be seen in Figures
18.1(b) and 18.1(c) respectively. The outermost circle of 18.1(b) depicts
unbounded thought (perhaps something like a universal Turing machine),
echoed in unbounded rationality in the outermost circle of 18.1(c) both
out of reach for humanitys nitary context. Bounded rationality, in contrast,
acknowledges universal processing constraints on human thought, just as
language universals give us access to processing constraints on human
language. Bounded rationality in the intermediate circle of 18.1(c) acknowledges an ought implies can sort of acknowledgement that thinking
universals, represented in the intermediate circle of 18.1(b), are shaped by
nitary processing constraints.
Bounded rationality neglects, however, the innermost circle of Figure
18.1(b), that of the epistemic context, exemplied in cultural diversity and
individual dierences. This is acknowledged in grounded rationality, in the
innermost circle of 18.1(c). As particular languages pick specic linguistic
tools from this limited space (Jackendo, 2002), so may particular epistemic
contexts combinations of cultures and individuals pick out specic thinking tools. For example, decontextualizing cultures tend to pick object-based
thinking, whereas holistic cultures tend to be, by denition, context based
(Buchtel & Norenzayan, 2009). Here, too, constraints often seem to be
soft, making leading authors refer to thinking styles (Nisbett et al., 2001;
Norenzayan, Smith, Kim, & Nisbett, 2002) rather than dierences in cognitive architecture (Buchtel & Norenzayan, 2009; Evans, 2008, 2009b).
Note that empirical evidence will always be gathered at the innermost of
the concentric circles, and generalized by comparing the space taken up by
dierent languages or dierent epistemic contexts. To borrow a distinction
from Harris (1980), we have no access to Language (with a capital L), only to
languages. With just one language, or one type of epistemic context (such as
culture) to draw on, overgeneralization is a clear hazard. Without crosslinguistic studies, it is easy to mistake particular language constraints for
universal constraints, a sin typically committed by traditional grammar.
Similarly, with no cross-cultural comparison, evidence from a specic culture
can easily be mistaken for thinking universals (see Norenzayan & Heine,
2005, for a similar argument). The point is that both linguistic universals and
thinking universals are a result of second-order theorizing. We only have
access to human cognition in particular epistemic contexts. Unlike bounded
rationality, grounded rationality takes this into account.4
Bounded and grounded rationality considerations alike are constraints
rather than positive theories. Neither can dictate what the content of rationality norms is; they can only determine what it cannot be: norms can be neither
computationally intractable nor contextually alien, respectively. The question
therefore remains what the content of relativist, epistemic-context sensitive

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

407

norms would actually be. A theory of the content of rational norms that is,
what dierent cultures consider to be norms of rational thinking would
have to be accessed descriptively. This is part of the task that the crosscultural approach undertook in recent years (e.g., Buchtel & Norenzayan,
2009; Choi et al., 2007; Koo & Choi, 2005; Nisbett et al., 2001).
If there is an answer, it has to be found in some level of abstraction, an
approach dominant in mainstream linguistics (Chomsky, 1965, pp. 2730;
Jackendo, 2002, pp. 7478) and voiced in psychology, for example, by
Norenzayan and Heine (2005). This is a hard-core approach to universals:
the idea that at some level, there lurks a formalized or natural universal which
is foundational to all particular manifestations. Such a hard core may constitute, in Cherniaks (1986) terms, an element of minimal rationality (see also
Saunders & Over, 2009). For psychological universals in general, the evidence
does seem to support a universal hard core. It is not obvious, however, that
such strong psychological universals, common to all epistemic contexts,
can also be found in the domain of rational norms. The evidence that
Norenzayan and Heine review only supports the existence of a hard core of
psychological tools; it does not conclusively support that rational norms,
specically, are among these hard-core universal tools. If there is a universal,
perhaps it is only the existence of norms simpliciter, as Carruthers (2006)
recently suggested not necessarily rational norms. An alternative relativist
view would be that such a universal hard core for rational norms will never be
found. Perhaps rational norms belong to the relativist part of the psychological toolbox. Or perhaps normative universals are like Wittgensteinian
families, with partially overlapping features but no universal hard core, whatever the level of abstraction (for a similar proposal see Stich, 1990, p. 88). For
example, perhaps there is a family of cognitive systems that shares noncontradiction as a rule, and another cognitive family that depends on dialectics;
while both families share, say, a correspondence conception of truth. Note
that this Wittgensteinian option is a possible alternative to any claim of
universalism, but would be rather dicult to determine empirically. A limited
survey of cognitive systems may come up with a feature that seems universal,
but may not be shared by another cognitive family, left out of the survey.
Universal constraints on rationality, then, still allow a wide margin for
local factors, both because nitary boundaries are somewhat elastic, and
because the question of specic solutions remains open. As pointed out earlier,
however, the link from is to ought is a particularly tenuous one. That
cognition varies in epistemic context does not necessarily mean that normative rationality should take this into account. In other words, it is perfectly
possible to acknowledge the innermost circle in Figure 18.1(b), and yet reject
its analogue in 18.1(c), acknowledging only the universal constraints in the
middle circle. The most extreme form of relativism would focus on grounded
rationality, the particular component, to the exclusion of bounded rationality, the universal component. I doubt, however, that such a position is viable.
It seems beyond dispute that the intermediate, universal circle should be

http://www.psypress.com/the-science-of-reason-9781848720152

408

Elqayam

acknowledged, and indeed bounded rationality considerations seem to be a

consensus in the rationality debate regardless of position. However, it is certainly possible to adopt a universalist position, ignoring grounded rationality
altogether. The question, then, is if grounded rationality constraints should
be taken into account at all, given their particular and local nature. The next
section examines this issue.

Weak versus strong normative relativism: The question

of universals
Given bounded rationality as a departure point, one can still accept grounded
rationality as an additional relativist set of constraints, or reject it. Weak
relativism would accept bounded rationality constraints but reject grounded
rationality, whereas strong relativism would acknowledge both types of constraints. To reiterate, the question is not whether bounded rationality, the
intermediate circle of Figure 18.1(c), should be accepted as a relevant set of
constraints on what accounts as normative: that much goes without saying.
Indeed in this weak sense, all contributors to the rationality debate can be
said to hold a relativist position. The question is whether grounded rationality, or constraints sensitive to epistemic context, should be ignored, while
bounded rationality, or more generally universal constraints on normative
rationality, should be given a privileged status. Nozick (1993), for example,
argues that if norms are particular to epistemic context place, time, or
individual then they are just accidental generalizations. It is only nomic
universals scientic lawlike statements that can provide viable norms.
In this section I will compare these weak and strong approaches to normative relativism. On the face of it, the idea that cognitive universals should have
a special place in our normative considerations, over and above that of contextual systems, seems an appealing and sensible position to take. Weak normative relativism acknowledges relative constraints recall that bounded
rationality itself is relativist to some degree while avoiding the potentially
unsettling implications of strong normative relativism. It even seems to
deect the criticism of epistemic chauvinism, since only universals are taken
into account. Suppose, for example, that a principle of noncontradiction is a
cognitive rational universal (as suggested by, e.g., Over, 2004). If all humans
have access to this principle, and some individuals violate it, no one can
accuse us of being epistemic chauvinists if we condemn them as irrational.
Weak normative relativism is also congruent with the amalgam solution to
the isought fallacy presented above. Even if dierent cultures construct different factual-evaluative compounds, the very fact that such compounds exist
seems to be universal.
It seems, then, that weak normative relativism carries all of relativisms
blessings and none of its hazards; that while we have to acknowledge
bounded rationality, we can safely ignore grounded rationality. However,
there are still good reasons to adopt grounded rationality in addition to the

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

409

bounded sort that is, good reasons to prefer strong relativism to the weak
version, as the next section proposes.
How normative universals vary
The most basic argument in favour of strong relativism, the one examined in
the beginning of this section, is the ought implies can principle. If there are
local constraints on cognitive mechanisms, then they have to be taken into
account when setting rational norms, alongside universal constraints: one can
hardly be accused of being irrational if one does not use cognitive tools that
one cannot access. It makes little dierence if a particular cognitive tool
cannot be accessed because it is computationally intractable, and therefore
inaccessible to all humans, or just not part of the cognitive repertoire of a
particular epistemic context, and accessible to some individuals but not
others. Cant is cant.
A possible objection here is that ought implies can may be a good reason
for bounded rationality, but not for grounded rationality, because the intermediate circle reects genuine processing constraints, whereas the innermost
circle reects mere optional constraints. In other words, there is a qualitative
dierence between bounded rationality constraints and grounded rationality
constraints: the former makes a much more powerful cant than the latter.
Training, for example, can transcend cultural constraints (Koo & Choi,
2005). Hence, if the rationale is that processing limitations have to be taken
into account, and if these limitations are much less severe for epistemic
context, then ought implies can can be used to justify bounded but not
grounded rationality.
However, as pointed out earlier in this section, constraints of bounded
rationality, too, may be less rigid than they seem. Recall that in natural language universals there may be exceptions to the rules at least in some cases. In
our example of word orders, despite the processing diculty, some examples
of each of the six possible word orders do exist in the world languages.
Granted some are very rare, one cannot defend a cannot thesis in any
absolute sense here. Even nitary boundaries vary across individuals. True,
no one lives for ever; nevertheless, nitary limits on computation are not the
same for everyone. There are always outliers with exceptional abilities, such as
Lurias famous mnemonic genius S (Luria, 2002), whose memory for long
lists remained intact decades after memorizing them.
The point is that cognitive universals come in various shapes and degrees.
Norenzayan and Heine (2005), for example, distinguish between accessibility
universals, functional universals, and existential universals. By their typology,
all universals share cognitive tools, but only accessibility universals use the
same cognitive tool for the same function and with the same degree of accessibility for all cultures (or, in the terminology suggested here, for all epistemic
contexts). Functional universals use the same cognitive tools for the same
cognitive function but with varying degrees of accessibility, and existential

http://www.psypress.com/the-science-of-reason-9781848720152

410

Elqayam

universals, the weakest type of universals, use the same cognitive tools for
dierent functions.
Even within bounded rationality, then, constraints are graded rather than
absolute. If a qualitative line is to be drawn, it might be easier to defend
within bounded rationality, than between bounded and grounded rationality.
For example, one may claim that only access universals are genuine cognitive
constraints that need to be taken into account in bounded rationality. The
risk of this approach is that one may be left with an empty set, a theoretical
notion that has intension but no extension. As Norenzayan and Heine point
out, robust empirical support for such hard-core universals can be rather
dicult to come by. There is always the possibility that some culture, some
language, some individual, would prove an exception by function or access.
The only viable alternative is to give up ought implies can altogether, which
seems radical and dicult to justify. To maintain ought implies can, then,
grounded rationality should be acknowledged.
The observation that normative constraints vary reects on another argument in favour of weak relativism, the epistemic chauvinism argument.
Recall that weak relativism seems to be no less successful than the strong sort
in deecting such accusations. However, since cognitive constraints vary in
their accessibility, epistemic chauvinism is still a non-negligible risk with
weak relativism. For example, decontextualizing tools seem to be existential
rather than access universals: they are universal in that they are shared by
all humans, but some individuals and cultures can access them more easily
then others. If we take decontextualization for a normative universal, we still
risk epistemic chauvinism towards individuals and cultures that prefer
contextualization.
In conclusion, I see little reason to prefer universal norms to culturally
sensitive ones. Both need to be acknowledged. This means that all rational
norms potentially take part in the normative game. For each individual in
each specic epistemic context, the relevant norms would include the ones
dictated by that epistemic context relevant to that individual, as well as
any universal norms. Normative rationality would still have to be strongly
relativistic, even while incorporating universal elements.

Grounded rationality, Panglossianism, and dual processing

With grounded rationality dened, we can now get back to the big rationality
debate. Are humans, then, rational? With this question in mind, grounded
rationality still seems a rather risky business. One particular pitfall is the
Panglossian (or anything goes) risk; the second and related one is loss of
relevance to cognitive architecture. The term Panglossian was coined
by Gould and Lewontin (1979), after Voltaires ever-optimistic protagonist
Dr Pangloss, and later on adopted by Stanovich (1999) in the context of the
rationality debate. (Ironically perhaps for our context, the name Pangloss
means all languages.) The Panglossian view maintains that humans are a

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

411

priori rational, a view introduced into the rationality arena by Cohen (1981),
and taken up by such diverse psychological programmes as rational analysis
(Oaksford & Chater, 1998, 2007) and the fast and frugal approach
(Gigerenzer & Selten, 2001; Gigerenzer & Todd, 1999).
The problem with grounded rationality seems to be, rst, that it invites a
you are right too type of Panglossianism; and, even worse, that it seems to
invite a particularly unattractive version it. The argument goes like this:
surely everyone follows some sort of normative rules, however harebrained.
Thus, if we allow room for relativism of this kind, then not only does everyone becomes rational, it is an uninteresting and trivial sort of rationality.
Stich (1990) calls this the anything goes argument: evaluation is no longer
possible because it is entirely subjugated to the particular. Oaksford and
Chater (2007) seem to have this type of the anything goes argument in
mind, when they argue that one can always claim rationality by cooking up a
bizarre set of assumptions about a problem that a person thinks they are
solving (Oaksford & Chater, 2007, p. 31). It seems, then, that if everyone is
rational in some context, this obviates the need for a good psychological
theory of rationality: the link between rationality and cognition is lost. The
challenge, then, is to create a framework for normative relativism that avoids
the anything goes pitfall. This is what the present section attempts to
achieve.
On its own, grounded rationality does not necessarily imply any specic
cognitive architecture. However, it is potentially compatible with a dualprocessing framework, and with this added set of constraints, the anything
goes argument loses much of its force. Dual-processing theories vary (for a
recent review see Evans, 2008), but they all share a distinction between one
family of processes, variably described as rapid, automatic, unconscious,
high capacity, and parallel, and another family of processes described
as slow, controlled, explicit, high eort, sequential, and correlated with
general ability (e.g., Evans, 2006, 2007; Evans & Over, 1996; Sloman, 1996;
Stanovich, 1999, 2004; Stanovich & West, 2000). Labels vary too, so here
I will adopt the neutral labels Type 1 and Type 2 respectively (Evans,
2009b; Frankish & Evans, 2009).
Dual-process theorists tend to conceptualize normative rationality as
linked to Type 2 processing, although again, theories vary in the particulars
of this connection. While no theory seems to support a one-to-one parallel,
what they do share is the idea that Type 2 processes are necessary (though not
sucient) for normative rationality. Thus, according to Evans and Over
(1996), one is rational2 when one explicitly conforms to a normative theory:
To possess rationality2, people need to have good reasons for what they are
doing, which must be part of an explanation of their action. They have to
follow rules sanctioned by a normative theory: this is what makes the reasons
good ones (p. 9). The crucial aspect here is the existence of explicit (hence
Type 2) normative rationale; good, which seems evaluative, means no more
and no less than normative. Stanovichs early conception of normative

http://www.psypress.com/the-science-of-reason-9781848720152

412

Elqayam

rationality is broader, emphasizing individuals and goals rather than processes (Stanovich, 1999, 2004), but he too, champions the crucial role of Type
2 processes in overriding the evolutionary goals that Type 1 processes are
adapted to. Later writings, both sets of authors emphasize that Type 2 processes can fail for example, Evans (2007) refers to a fundamental analytic
bias, which Stanovich (2009) calls serial associative cognition but their
earlier approaches viewed them as necessary for normative rationality (as
opposed to mere normative responding).
Of these two approaches, Stanovichs seems to be dicult to reconcile with
grounded rationality, since his approach is explicitly broad that is, it
encompasses critique of beliefs and desires (Stanovich, 2004). In contrast,
although Evans and Overs (1996) rationality2 is silent on the issue of culturesensitive norms, there seems to be nothing to preclude conformity to such
norms from being rational2, so long as this conformity is made explicit. The
advantage of this approach is that it manages to avoid both Panglossian and
epistemic chauvinism concerns at the same time. Let me examine what this
proposal entails. In a chapter for Jonathans Festschrift book, you just cant
do without an Arsenal example. So here is one:
Joe and Tom, an engineer and accountant respectively, and ardent
Arsenal fans both, discuss the latest successful match. The outcome was
a surprise: Arsenal has won despite the absence of a key player due to
injury the week before. I knew all along they would win, says Joe. Tom
agrees: They just had to, they have been on a winning streak since the
season started, he adds.
Thousands of miles away, Haneul and Chul, students of Oriental
medicine in a prestigious Korean university (Koo & Choi, 2005) and avid
followers of English football, discuss the same match: why has Bendtner
seemed out of form? Chul feels that there must be many complex reasons.
Haneul agrees. The wise man does not judge or divide, she quotes from
the Tao Te Ching (Lao Tzu, 2001, p. 102).
All four people contextualize, but they vary in explicit conformity to cultural
norms. Haneul and Chul both conform to holistic cultural norms by weighing complex contextual causes to performance.5 However, only Haneul has an
explicit reason for her judgement. By explicitly following Taoism, Haneul
conforms to a system that is normative for her particular epistemic context.
This context includes her holistic culture as well as her training in Oriental
medicine (recall that responses of psychology students in a comparably prestigious Korean university were indistinguishable from those of Western participants; Koo & Choi, 2005). By this view, then, Haneul is rational2, but Chul,
who follows the same rules implicitly and with no apparent reection, is not.
Haneul is no less rational2 than, say, a participant from a decontextualizing
culture taking part in a conjunction fallacy (Tversky & Kahneman, 1983)

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

413

experiment, who estimates that Linda is more likely to be a banker than a

feminist banker, and gives an explicit probabilistic explanation.
Now compare Haneul and Chul with Joe and Tom. The latter two fall prey
to hindsight (I knew it all along) bias, which falls outside the decontextualizing rational norms of their professions. However, Joe does so implicitly,
whereas Tom has an explicit rationale. The rationale, however, is not good:
a version of the hot hand fallacy, it does not conform to a culturally
relevant expert norm. Explicit but non-normative, Toms explanation falls
into the category of what Stanovich recently called contaminated mindware (Stanovich, 2009). This means that neither Joe nor Tom is rational2.
In an open peer commentary discussion of Rationality and Reasoning
(Evans & Over, 1997; Over & Evans, 1997), Evans and Over claried that they
considered rationality1 as basic: rationality2 is justied by providing good
means of achieving goals (Over & Evans, 1997, pp. 255256). Prima facie,
it seems that for Haneul, this is indeed the case: by explicitly conforming to
the principles of the Tao Te Ching, she can achieve everyday goals, such as
success in her chosen profession and esteem of her peers. She can even
achieve her epistemic goals; spiritual harmony, for example.
Evans and Overs suggestion, that normative rationality has to include
explicit normative reasons, is rather attractive, and as we just saw, perfectly
compatible with a relativist solution, so long as we include sensitivity to
epistemic context as a caveat. The main attraction is that it allows us to
distinguish between Haneul, who is rational2, and, say, a tree, which may
be in perfect spiritual harmony with the universe but cannot be said to be
normatively rational. This is an extension of the original suggestion, which
distinguished between the rational1 behaviour of a bumble-bee and consequentialist, rational2 decision making (Evans & Over, 1996, p. 148). Implicit
rule conformity does not make one rational2, and this applies to spiritual
harmony just as it applies to consequential decision making. Explicit rule
following does so, and again this applies just as well to spiritual harmony. The
fact that some of us may strongly reject spiritual harmony as an epistemic
goal is entirely beside the point, if we opt for normative relativism.
Note that this solution obviates Panglossianism worries. One can fail to be
normatively rational by failing to conform explicitly to a normative system;
the addition is that this normative system has to be relevant to the individual
within her or his epistemic context (in our example, this is a combination of
professional and geographical context). Thus, Haneul and Tom both respond
contextually and explicitly, but only Haneul is normatively rational, since
only she conforms to an epistemically relevant system.
Two caveats are in order. One is that normative responses should not be
taken as indication for explicit processing what Elqayam and Evans dubbed
oughtis fallacy. This is still true for the relativist version of normative
rationality. The second, and perhaps more interesting caveat, is that what falls
under the heading of explicit processing can vary considerably: it would be
analytic for decontextualizing epistemic contexts, but it may well be holistic

http://www.psypress.com/the-science-of-reason-9781848720152

414

Elqayam

for contextualizing cultures and individuals (Buchtel & Norenzayan, 2009).

This is a strength of grounded rationality. Consider the case of fundamental
attribution error (correspondence bias), the tendency to attribute other
peoples behaviours to their disposition and personality, discounting the
eects of context. People in decontextualizing cultures are much more likely
to commit this particular error, since they focus on the object rather than
its environment (e.g., Choi, Nisbett, & Norenzayan, 1999). Decontextualization in itself, then, does not guarantee inoculation against bias, in the
conventional absolutist sense used by psychologists.
In conclusion, a dual-processing framework of rationality does not necessarily acknowledge grounded rationality concerns; indeed, the broad version
advocated by Stanovich seems to contradict it. However, processing-based
versions such as Evans and Overs seem to be potentially compatible. From
a grounded rationality point of view, the advantage in adopting a dualprocessing framework seems to be a more rounded and psychologically
relevant perspective, and avoiding the unpalatable anything goes version
of the Panglossian position, that humans are necessarily and unfailingly
rational.

Grounded rationality and the descriptive agenda

Before I wind up this discussion, there is one last thing that needs to be
examined, and that is the relationship between the grounded rationality
framework presented in this chapter, and the descriptivist agenda defended
by Elqayam and Evans (2010). A possible objection to grounded rationality is
that it has an implicit descriptivist agenda: that it pulls the carpet from
underneath normative rationality. With a relativist framework, what constitutes normative rationality is no longer an ought question. It becomes a
purely descriptive question: the question of which normative system holds for
a particular agent in a particular epistemic context, and to what extent they
satisfy it. Ultimately, then, there is no such thing as normative relativism. The
ought is replaced with an is.
There are two possible answers here. One is that, even with a relativist
framework, evaluation is still possible. As we saw in the previous section, even
with a grounded rationality framework, people can still fail to be normatively
rational in a variety of nontrivial ways. Stich (1990) defended a similar argument, proposing that agents can still fail to be rational within a relativist
context, if they fail to adhere to their own standards. Hence, even with a
relativist framework of the sort inherent in grounded rationality, normative
evaluation is still attainable to those who want it.
The other answer, however, is precisely the opposite, and that is the answer
that I tend to favour, at least as far as the empirical study of human thought
is concerned. I do not deny that relativism does weaken the ought in normative rationality, and ultimately replaces it with an is question. However,
such an outcome, far from being a detriment, is actually quite desirable for

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

415

psychology. Elqayam and Evans (2010) conclude with an outline a descriptivist agenda for the psychology of reasoning and judgement and decision
making. Such a framework, they argue, proved benecial to the development
of linguistics as a mature science, and may carry similar benets for
psychology. The same observation is still relevant here.
In conclusion, while grounded rationality is an anathema to the absolutist
concerns of normativism, it still allows some scope for normative research
questions. Within empirical psychological science, grounded rationality sits
particularly well with a descriptivist agenda for the psychology of human
thought. By now, I hope I have convinced you at least that normative relativism, in the form of grounded rationality, should be taken seriously, and that it
can provide an interesting and productive framework for normative rationality and the associated empirical paradigms of human thought. Perhaps not
everyone is right, but what is right and what is wrong cannot be taken out of
context.

Notes
1 I will not address the related though separable issue of moral relativism (for a recent
review see Gowans, 2008), although I will draw on it occasionally.
2 The term is inspired by (rather than directly draws on) Barsalous (2008) notion of
grounded cognition.
3 By cross-cultural I mean any study that systematically compares two cultures or
more. This covers what Norenzayan and Heine (2005) call two-culture and threeculture or triangulation studies, as well as their genuine cross-cultural studies which
cover multiple cultures.
4 Some readers may be wondering where the linguistic equivalent of Figure 18.1(c)
has disappeared to. That is, if 18.1(b) is descriptive of human thought, and 18.1(c)
is prescriptive, why does Figure 18.1 include no representation of normative linguistic considerations, only the descriptive ones represented in 18.1(a)? The answer
is simple: Modern linguistic theory is concerned very little with normative considerations, a tradition that goes back as far as de Saussure (1959).
5 By doing so they avoid the fundamental attribution error. They have not, for
example, maintained that Bendtner always underperforms when he is under
pressure.

Acknowledgements
The research reported in this chapter was partially supported by research
travel funding from De Montfort University to the author. A preliminary
sketch of the arguments in this chapter was presented at the London Workshop on Reasoning in honour of Jonathan Evanss 60th birthday, held in
Birkbeck College in August 2008. I thank Jonathan Evans for detailed comments on previous versions. Lastly, my heartfelt thanks, as always, go to
Jonathan Evans, for inspiration, for mentorship, for friendship, for challenges
and disputes; and for gallons of coee.

http://www.psypress.com/the-science-of-reason-9781848720152

416

Elqayam

References
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59,
617645.
Buchtel, E., & Norenzayan, A. (2009). Thinking across cultures: Implications for dual
processes. In J. St B. T. Evans & K. Frankish (Eds.), In two minds: Dual processes
and beyond (pp. 217238). Oxford, UK: Oxford University Press.
Carruthers, P. (2006). The architecture of the mind. Oxford, UK: Oxford University Press.
Cherniak, C. (1986). Minimal rationality. Cambridge, MA: MIT Press.
Choi, I., Koo, M., & Choi, J. A. (2007). Individual dierences in analytic versus
holistic thinking. Personality and Social Psychology Bulletin, 33, 691705.
Choi, I., Nisbett, R. E., & Norenzayan, A. (1999). Causal attribution across cultures:
Variation and universality. Psychological Bulletin, 125, 4763.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Cohen, L. J. (1981). Can human irrationality be experimentally demonstrated?
Behavioral and Brain Sciences, 4, 317370.
Comrie, B. (1981). Language universals and linguistic typology: Syntax and morphology. Oxford, UK: Blackwell.
de Saussure, F. (1959). Course in general linguistics. New York: McGraw-Hill.
(Original work published 1916)
Elqayam, S. (2007). Normative rationality and the isought fallacy. In S. Vosniadou,
D. Kayser, & A. Protopapas (Eds.), Proceedings of the second European cognitive
science conference (pp. 294299). Hove, UK: Lawrence Erlbaum Associates Ltd.
Elqayam, S., & Evans, J. St B. T. (2010). Substracting ought from is: Descriptivism
versus normativism in the study of human thinking. Unpublished manuscript, De
Montfort University.
Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious.
American Psychologist, 49, 709724.
Epstein, S., Pacini, R., Denes-Raj, V., & Heier, H. (1996). Individual dierences in
intuitive-experiential and analytical-rational thinking styles. Journal of Personality
and Social Psychology, 71, 390405.
Evans, J. St B. T. (1993). Bias and rationality. In K. I. Manktelow & D. E. Over (Eds.),
Rationality: Psychological and philosophical perspectives (pp. 630). London:
Routledge.
Evans, J. St B. T. (2006). The heuristicanalytic theory of reasoning: Extension and
evaluation. Psychonomic Bulletin & Review, 13, 378395.
Evans, J. St B. T. (2007). Hypothetical thinking: Dual processes in reasoning and judgement. Hove, UK: Psychology Press.
Evans, J. St B. T. (2008). Dual-processing accounts of reasoning, judgment, and
social cognition. Annual Review of Psychology, 59, 255278.
Evans, J. St B. T. (2009a). Does rational analysis stand up to rational analysis?
Behavioral and Brain Sciences, 32, 8889.
Evans, J. St B. T. (2009b). How many dual process theories do we need: One, two or
many? In J. St B. T. Evans & K. Frankish (Eds.), In two minds: Dual processes and
beyond (pp. 3354). Oxford, UK: Oxford University Press.
Evans, J. St B. T., & Over, D. E. (1996). Rationality and reasoning. Hove, UK:
Psychology Press.
Evans, J. St B. T., & Over, D. E. (1997). Rationality in reasoning: The case of deductive
competence. Current Psychology of Cognition, 16, 338.

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

417

Evans, N., & Levinson, S. (2009). The myth of language universals: Language diversity
and its importance for cognitive science. Behavioral and Brain Sciences, 32, 429448.
Frankena, W. (1939). The naturalistic fallacy. Mind, 48, 464477.
Frankish, K., & Evans, J. St B. T. (2009). The duality of mind: A historical perspective. In J. St B. T. Evans & K. Frankish (Eds.), In two minds: Dual processes and
beyond. Oxford, UK: Oxford University Press.
Gigerenzer, G., & Selten, R. (2001). Bounded rationality: The adaptive toolbox.
Cambridge, MA: MIT Press.
Gigerenzer, G., & Todd, P. M. (1999). Simple heuristics that make us smart. Oxford,
UK: Oxford University Press.
Gottwald, S. (2001). A treatise on many-valued logics (Studies in logic and computation, Vol. 9). Baldock, UK: Research Studies Press Ltd.
Gould, S. J., & Lewontin, R. C. (1979). The spandrels of San Marco and the
Panglossian paradigm: A critique of the adaptationist programme. Proceedings of
the Royal Society of London, Series B, 205, 581598.
Gowans, C. (2008). Moral relativism. In E. N. Zalta (Ed.), Stanford encyclopedia
of philosophy (Winter 2008 edition). Stanford, CA: Stanford University. Retrieved
February 28, 2010, from http://plato.stanford.edu/entries/moral-relativism/
Greenberg, J. H. (1963). Some universals of grammar with particular reference to
the order of meaningful elements. In J. H. Greenberg (Ed.), Universals of language
(pp. 5890). Cambridge, MA: MIT Press.
Greenberg, M. J. (2008). Euclidean and non-Euclidean geometries (4th ed.). New York:
W. H. Freeman.
Harris, R. (1980). The language-makers. London: Duckworth.
Hawkins, J. A. (1983). Word order universals. New York: Academic Press.
Hofstadter, R. (1955). Social Darwinism in American thought. Boston, MA: Beacon
Press.
Hudson, W. D. (1969). The isought question: A collection of papers on the central
problem in moral philosophy. London: Macmillan.
Hume, D. (2000). A treatise on human nature. Oxford, UK: Clarendon Press. (Original
work published 17391740)
Jackendo, R. (2002). Foundations of language: Brain, meaning, grammar, evolution.
Oxford, UK: Oxford University Press.
Kant, I. (1932). Critique of pure reason. London: Macmillan. (Original work
published 1787)
Koo, M., & Choi, I. (2005). Becoming a holistic thinker: Training eect of oriental medicine on reasoning. Personality and Social Psychology Bulletin, 31,
12641272.
Lao Tzu (2001). The way of life: Dao de jing (R. B. Blakney, trans.). New York: New
American Library.
Lloyd, G. (1996). Science in antiquity: The Greek and Chinese cases and their
relevance to the problems of culture and cognition. In D. R. Olson & N. Torrance
(Eds.), Modes of thought: Explorations in culture and cognition (pp. 1533).
Cambridge, UK: Cambridge University Press.
Luria, A. R. (2002). The mind of a mnemonist: A little book about a vast memory.
Cambridge, UK: Harvard University Press.
Moore, G. E. (1903). Principia ethica. New York: Cambridge University Press.
Nichols, S. (2004). Folk concepts and intuitions: From philosophy to cognitive
science. Trends in Cognitive Sciences, 8, 514518.

http://www.psypress.com/the-science-of-reason-9781848720152

418

Elqayam

Nickerson, R. S. (2008). Aspects of rationality: Reections on what it means to be

rational and whether we are. New York: Psychology Press.
Nisbett, R., Peng, K., Choi, I., & Norenzayan, A. (2001). Culture and systems of
thought: Holistic vs analytic cognition. Psychological Review, 108, 291310.
Norenzayan, A., & Heine, S. J. (2005). Psychological universals: What are they and
how can we know? Psychological Bulletin, 131, 763784.
Norenzayan, A., Smith, E. E., Kim, B. J., & Nisbett, R. E. (2002). Cultural preferences
for formal versus intuitive reasoning. Cognitive Science, 26, 653684.
Nozick, R. (1993). The nature of rationality. Princeton, NJ: Princeton University
Press.
Oaksford, M., & Chater, N. (1998). Rationality in an uncertain world. Hove, UK:
Psychology Press.
Oaksford, M., & Chater, N. (2007). Bayesian rationality: The probabilistic approach to
human reasoning. Oxford, UK: Oxford University Press.
Oaksford, M., & Chater, N. (2009). The uncertain reasoner: Bayes, logic, and rationality. Behavioral and Brain Sciences, 32, 69120.
Olson, D. R., & Torrance, N. (1996). Modes of thought: Explorations in culture and
cognition. Cambridge, UK: Cambridge University Press.
Over, D. E. (2004). Rationality and the normative/descriptive distinction. In D. J.
Koehler & N. Harvey (Eds.), Blackwell handbook of judgment and decision making
(pp. 318). Oxford, UK: Blackwell.
Over, D. E., & Evans, J. St B. T. (1997). Two cheers for deductive competence. Current
Psychology of Cognition, 16, 255278.
Rescher, N. (1969). Many-valued logics. New York: McGraw-Hill.
Saunders, C. E., & Over, D. E. (2009). In two minds about rationality? In J. St B. T.
Evans & K. Frankish (Eds.), In two minds: Dual processes and beyond (pp. 317334).
Oxford, UK: Oxford University Press.
Schroyens, W. (2009). On is and ought: Levels of analysis and the descriptive versus
normative analysis of human reasoning. Behavioral and Brain Sciences, 32,
101102.
Schurz, G. (1997). The isought problem: An investigation in philosophical logic.
Dordrecht: Kluwer.
Searle, J. R. (1964). How to derive ought from is. Philosophical Review,
73, 4358.
Simon, H. A. (1957). Models of man: Social and rational. New York: Wiley.
Simon, H. A. (1982). Models of bounded rationality. Cambridge, MA: MIT Press.
Simon, H. A. (1983). Reason in human aairs. Stanford, CA: Stanford University
Press.
Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological
Bulletin, 119, 322.
Stanovich, K. E. (1999). Who is rational? Studies of individual dierences in
reasoning. Mahwah, NJ: Lawrence Elrbaum Associates, Inc.
Stanovich, K. E. (2004). The robots rebellion: Finding meaning in the age of Darwin.
Chicago: Chicago University Press.
Stanovich, K. E. (2009). Distinguishing the reective, algorithmic, and autonomous
minds: Is it time for a tri-process theory? In J. St B. T. Evans & K. Frankish (Eds.),
In two minds: Dual processes and beyond. Oxford, UK: Oxford University Press.
Stanovich, K. E., & West, R. F. (2000). Individual dierences in reasoning: Implications for the rationality debate. Behavioral and Brain Sciences, 23, 645726.

http://www.psypress.com/the-science-of-reason-9781848720152

Grounded rationality

419

Stein, E. (1996). Without good reason: The rationality debate in philosophy and cognitive science. Oxford, UK: Oxford University Press.
Stich, S. P. (1990). The fragmentation of reason: Preface to a pragmatic theory of
cognitive evaluation. Cambridge, MA: MIT Press.
Tversky, A., & Kahneman, D. (1983). Extensional vs intuitive reasoning: The conjunction fallacy in probability judgment. Psychological Review, 90, 293315.
Weinberg, J. M., Nichols, S., & Stich, S. P. (2006). Normativity and epistemic
intuitions. In R.Viale, D. Andler, & L. A. Hirschfeld (Eds.), Biological and cultural
bases of human inference (pp. 191222). Mahwah, NJ: Lawrence Erlbaum
Associates, Inc.
Weinstein, D. (2002). Herbert Spencer. In E. N. Zalta (Ed.), Stanford encyclopedia
of philosophy (Winter 2002 edition). Stanford, CA: Stanford University. Retrieved
February 28, 2010 from http://plato.stanford.edu/entries/spencer/
Williams, B. (1985). Ethics and the limits of philosophy. Cambridge, MA: Harvard
University Press.
Yama, H., Nishioka, M., Horishita, T., Kawasaki, Y., & Taniguchi, J. (2007). A dual
process model for cultural dierences in thought. Mind and Society, 6, 143172.

http://www.psypress.com/the-science-of-reason-9781848720152